⚠️⚠️⚠️⚠️⚠️⚠️Important Notes Before Use⚠️⚠️⚠️⚠️⚠️⚠️

Greatly improve!

Thanks to v1.0, I found out where the mistakes were and retrained it once.

The v1.0 version is a bit overtrainned,resulting in very poor image generation quality.

Compared to v1.0, v2.0 version has been greatly improved. I’m really happy — this should be my most successful LoRA model.

Prompt:toshi style (required), uncensored(required),old style (optional, will cause some variations),

Weight:0.8 (recommended) ~ 0.9

⚠️⚠️⚠️⚠️⚠️⚠️用前须知⚠️⚠️⚠️⚠️⚠️

得益于v1.0，我知道了错误的地方并重新训练了一遍

v1.0版本有点过拟合了,导致不是原图训练的tag，出图效果很差

相较于v1.0有了巨大提升，我非常开心，这应该是我最成功的lora模型了

提示词: toshi style(必须),uncensored(必须),old style(可选，会有点变化),

权重:0.8(推荐)~0.9

(say in a low voice)

If you enjoy it, I’d really appreciate your support😭, such as buying me a coffee!

如果你喜欢它，真的很感谢你的支持😭，比如请我喝杯咖啡吧！